Process of Selecting Portfolio Managers Based on Automated Artificial Intelligence Techniques

1. A computerized system encoded with a method of qualifying a portfolio management firm, the method comprising the steps of: (a) calculation of an at least three dimensional multi-dimensional space of points associated with characteristics of firms, the calculation using a neural network characterized by synaptic neuron weightings; (b) evaluation and processing of the multi-dimensional space including projecting the points in the at least three-dimensional multi-dimensional space onto a 2-dimensional space using a Kohonen network, the projecting creating a feature map, wherein firms placed on this map are described by the characteristics of the regions in which they appear on the map, thereby helping visualize the attributes of the firms; (c) detection of groups corresponding to firms having similar characteristics in the feature map; (d) labelling each group; and applying selection criteria to qualify a firm suitable for a particular mandate, wherein the method further includes a step of transforming unstructured raw data of an RFP into a multi-dimensional space of metrics permitting a rational selection of a portfolio management firm.

2. A computerized system encoded with a method of qualifying a portfolio management firm, the method comprising the steps of: (a) calculation of an at least three dimensional multi-dimensional space of points associated with characteristics of firms, the calculation using a neural network characterized by synaptic neuron weightings; (b) evaluation and processing of the multi-dimensional space including projecting the points in the at least three-dimensional multi-dimensional space onto a 2-dimensional space using a Kohonen network, the projecting creating a feature map, wherein firms placed on this map are described by the characteristics of the regions in which they appear on the map, thereby helping visualize the attributes of the firms; (c) detection of groups corresponding to firms having similar characteristics in the feature map; (d) labelling each group; and applying selection criteria to qualify a firm suitable for a particular mandate, wherein the method includes adaption routines that adapt the method according to a property of a neural network known as plasticity wherein synaptic weightings in the neural network demonstrate learning abilities.

3. The system of claim 1 wherein firm characteristics for calculation of the multi-dimensional space are responses by designates of firms to questions in a request for proposals for a given mandate, such responses being used to perform a calculation applying artificial intelligence and data mining techniques, the calculation resulting in a comparison of one firm with another such that a sponsor is able to more easily discern differences between offerings of one firm with another, according to requirements of his mandate.

4. The system of claim 3 wherein firm characteristics for calculation of the multi-dimensional space are responses by designates of firms to questions in a request for proposals for a given mandate, such responses being used to perform a calculation applying artificial intelligence and data mining techniques, the calculation resulting in a comparison of one firm with another such that a sponsor is able to more easily discern differences between offerings of one firm with another, according to requirements of his mandate.

5. The system f claim 3 wherein the points in the multi-dimensional space are projected onto a 2-dimensional space to better visualize the attributes of the firms.

6. The system of claim 5 wherein the projecting is based upon a Principal Component Analysis technique.

7. The system of claim 5 , wherein the multi-dimensional space is a ten-dimensional space comprising ten values between 0 and 1, each value representing a dimension, wherein the higher the value of the dimension, the greater the correlation to a concept corresponding to the dimension.

8. The system of claim 7 , wherein the ten-dimensional space is generated by a connection scheme and synaptic vectors of the neural network operating on responses to questions in a request for proposal.

9. The system of claim 3 wherein the method operates on a network selected from a group of networks consisting of a virtual private network, a LAN, and a distributed network known as the Internet.

10. The system of claim 1 , wherein the method further includes a step of transforming-unstructured raw data of an RFP into the multi-dimensional space.

11. The system of claim 1 , wherein the method includes adaption routines that adapt the method according to a property of a neural network known as plasticity wherein synaptic weightings in the neural network demonstrate learning abilities.

12. The system of claim 2 , wherein the method operates on a network selected from a group of networks consisting of a virtual private network, a LAN, and a distributed network known as the Internet.

13. A service product transmittable over a network, the product being a recommendation or set of recommendations of portfolio management firms, the recommendation being obtained by applying the system of claim 1 .

14. The service product of claim 13 , wherein the network is a network selected from a group of networks consisting of a virtual private network, a LAN, and a distributed network known as the Internet.

15. A computerized system encoded with a method of qualifying a portfolio management firm, the method comprising the steps of: (a) calculation of an at least three dimensional multi-dimensional space of points associated with characteristics of firms, the calculation using a neural network characterized by synaptic neuron weightings; (b) evaluation and processing of the multi-dimensional space including projecting the points in the at least three-dimensional multi-dimensional space onto a 2-dimensional space using a Kohonen network, the projecting creating a feature map, wherein firms placed on this map are described by the characteristics of the regions in which they appear on the map, thereby helping visualize the attributes of the firms; (c) detection of groups corresponding to firms having similar characteristics in the feature map; (d) labelling each group; and applying selection criteria to qualify a firm suitable for a particular mandate, wherein, prior to calculation of a multi-dimensional space, inputs are gathered using an online questionnaire consisting of questions which are designed for automated analysis.

16. The system of claim 15 , wherein the questions comprise question types selected from a group of question types consisting of atomic, vector, and matrix questions.

17. The system of claim 16 , wherein answers input to the questions are savable for later retrieval by the firm being queried.